TSDZ2 mid drive with 860C, 850C or SW102 displays only -- Flexible OpenSource firmware (Casainho code only)

[eyebyesickle]

... I yes, I am also buying the Blue gear for 20 euros on Aliexpress straight to my door, the only thing is that it takes some time and hand dirty to replace it inside the motor

Do you have a good reliable source on Aliexpress please? There are so many, and it's good to know the good from the bad. Thanks

If anyone here on ES (active users only, no new posters) needs blue gears, send me a PM with an email, and I can send you a link, I will send them for free, you only choose shipping and pay postage. I literally have a bin full of them and happy to share with the community

 

[feyris]

https://github.com/OpenSource-EBike-firmware/TSDZ2_wiki/wiki/Features-and-configurations-on-display

UP and DOWN long press to select the variable
ON/OFF long press to leave
UP or DOWN click to choose a new variable
ON/OFF click to jump to next field


Go to the second screen and keep up AND down key together pressed for sometime

Great thanks! I missed the 'AND' part :thumb:

Does anyone know where I can get the init packet file for the SW102 firmware update from?

Both nRF Toolbox and nRF Connect (both from the play store) ask for this file now and won't download the .hex file to the SW102 over Bluetooth without it.

Using NrfToolbox 2.8.4 from the playstore, I didn't need any other file than the .hex

 

[HughF]

https://github.com/OpenSource-EBike-firmware/TSDZ2_wiki/wiki/Features-and-configurations-on-display

UP and DOWN long press to select the variable
ON/OFF long press to leave
UP or DOWN click to choose a new variable
ON/OFF click to jump to next field


Go to the second screen and keep up AND down key together pressed for sometime

Great thanks! I missed the 'AND' part :thumb:

Does anyone know where I can get the init packet file for the SW102 firmware update from?

Both nRF Toolbox and nRF Connect (both from the play store) ask for this file now and won't download the .hex file to the SW102 over Bluetooth without it.

Using NrfToolbox 2.8.4 from the playstore, I didn't need any other file than the .hex

Interesting, using that same version of nrfToolbox from the play store, it refused to flash my SW102 with the .hex file due to the complaints about the lack of an init packet file. I've just got a new phone so I'll try it again - the bluetooth was always flaky on my previous handset

 

[skestans]

... I yes, I am also buying the Blue gear for 20 euros on Aliexpress straight to my door, the only thing is that it takes some time and hand dirty to replace it inside the motor

Do you have a good reliable source on Aliexpress please? There are so many, and it's good to know the good from the bad. Thanks

If anyone here on ES (active users only, no new posters) needs blue gears, send me a PM with an email, and I can send you a link, I will send them for free, you only choose shipping and pay postage. I literally have a bin full of them and happy to share with the community

Nice of you, thanks!

 

[izeman]

... I yes, I am also buying the Blue gear for 20 euros on Aliexpress straight to my door, the only thing is that it takes some time and hand dirty to replace it inside the motor

Do you have a good reliable source on Aliexpress please? There are so many, and it's good to know the good from the bad. Thanks

If anyone here on ES (active users only, no new posters) needs blue gears, send me a PM with an email, and I can send you a link, I will send them for free, you only choose shipping and pay postage. I literally have a bin full of them and happy to share with the community

Ordered one, and it's already on the way. Thanks to eyebyesickle for the generous offer!!! 8) :thumb:

 

[izeman]

Had a short maiden voyage today. All wires still hanging around, and no rear brake installed, but at least it moves. Even with the quite bad chain line all works fine. I'm still consider installing an IGH for a cleaner look.
Power wise it's REALLY a pedelec. It supports nicely, and smoothly but it still moves like a bicycle.
I also ordered some 3mm thick silicone pads to try the cooling mod.

One thing that I noticed instantly, especially with the short cranks (145mm): You reach 90rpm cadence in a instance. I don't get why motor power is cut off at 90rpm. I saw that mentioned somewhere, but why is it?
I could pedal much faster, but won't as I get no more power.

 

[baloubic]

I just finished my first ebike with the latest open source firmware (as of today).

Unfortunately I have the same issue as described a few pages ago, where I get no assist if my lights are on. Everything works great with lights off. When I do try to pedal with lights on, the 850C shows like 36amps current draw too (without motor activity).

Any updates as to what cause this? Did I do something wrong?

 

[feyris]

I just finished my first ebike with the latest open source firmware (as of today).

Unfortunately I have the same issue as described a few pages ago, where I get no assist if my lights are on. Everything works great with lights off. When I do try to pedal with lights on, the 850C shows like 36amps current draw too (without motor activity).

Any updates as to what cause this? Did I do something wrong?

I saw on gitHub that the light issue is on the pull requests.

Anyway, removed my lights from the tsdz2.
They now use a separate DC-DC converter on the main battery and are always on.
I can use much more power than before

 

[baloubic]

I see. Too bad, was awesome to switch with the display. Wonder if a small 6v relay would work better, allowing you to pass the signal from the display even though you draw power to the light from the battery?

Luckily, my light is variable voltage and has an on off switch, so wire straight to battery without buck converter is a no fuzz option.

 

[lvx_xvl]

Man that sucks. Are you using a speed sensor connector, or DuPont wires with heat shrink tube at the ends? If DuPont, maybe you’re not getting a good connection and should try with a speed sensor extension cable so you get the proper connector and good contact.

If your speed sensing works correctly (ie displays the right speed on the display) and you’ve tried so many computers and dongles, the only option left is a bad connection at the plug.

You could try with a Linux live disc and see if stm8flash works there but if it didn’t work on OS X...

I'm using the speedsensor connector with the Dupont wires (4cm long) and heat shrink tubes. And you are right that might be the issue, but they really snuck in tight. So I'm assuming (very bad I know) they work. If I connect the GND, 5V and SWIM and boot the bike via the VLCD5 display then my ST Link V2 boots, so I'm fairly confident (I know very bad) that the connection is working.

I used a speed sensor extension cable at first, but a lot of people point out to the long leads being the problem.

regards,
LVX

I took the motor controller board out today. It did come out easily.


I hooked up with the short Dupont wires (which still doesn't make sense, as the wire from the board up to connector is ~30cm long) and it did not connect. I tried 3 different type of STLink V2 dongles.

I removed some of the coating to measure from the PCB pads. All of the wires gave a good readout. I tried testing what is what, and with some help of google I ended up with (right-to-left):
brown = VCC
black = SWIM
orange = GND
purple = reset


Could it be that this board is somehow protected for readout, or is a protected firmware still communicating?

 

[feyris]

I see. Too bad, was awesome to switch with the display. Wonder if a small 6v relay would work better, allowing you to pass the signal from the display even though you draw power to the light from the battery?

Luckily, my light is variable voltage and has an on off switch, so wire straight to battery without buck converter is a no fuzz option.

I'm pretty sure that a 6v relay on the light signal will be ok.

 

[lvx_xvl]

Man that sucks. Are you using a speed sensor connector, or DuPont wires with heat shrink tube at the ends? If DuPont, maybe you’re not getting a good connection and should try with a speed sensor extension cable so you get the proper connector and good contact.

If your speed sensing works correctly (ie displays the right speed on the display) and you’ve tried so many computers and dongles, the only option left is a bad connection at the plug.

You could try with a Linux live disc and see if stm8flash works there but if it didn’t work on OS X...

I'm using the speedsensor connector with the Dupont wires (4cm long) and heat shrink tubes. And you are right that might be the issue, but they really snuck in tight. So I'm assuming (very bad I know) they work. If I connect the GND, 5V and SWIM and boot the bike via the VLCD5 display then my ST Link V2 boots, so I'm fairly confident (I know very bad) that the connection is working.

I used a speed sensor extension cable at first, but a lot of people point out to the long leads being the problem.

regards,
LVX

I took the motor controller board out today. It did come out easily.


I hooked up with the short Dupont wires (which still doesn't make sense, as the wire from the board up to connector is ~30cm long) and it did not connect. I tried 3 different type of STLink V2 dongles.

I removed some of the coating to measure from the PCB pads. All of the wires gave a good readout. I tried testing what is what, and with some help of google I ended up with (right-to-left):
brown = VCC
black = SWIM
orange = GND
purple = reset


Could it be that this board is somehow protected for readout, or is a protected firmware still communicating?

I just wanted to add that I was doing some debugging to see if I could find out if the board is actually powered by the ST Link.

So I hooked up an UART dongle to read out the main connector, the 9 pin version. Immediately after plugging in the ST Link dongle the UART started to write data to the console, all 16 byte lines of code.


As soon as I click on the STVP "read all tabs" icon, the UART stopped. So it definitely rebooted the MCU.

I then went on to check if I have to correct MCU, which luckily I did.


Can someone tell me if can try out a different type of test to see if the SWIM works?

regards,
LVX

 

[hefest]

Ok, so if assist level is set too high you get less assist. I've set highest assist level to 0.9 and couldn't get more than 450-500W up a steep hill. Reduced to 0.200 motor is assisting now with 650W max, still not 750W I was getting with 0.19 firmware.

Another really strange thing is that if I really push pedals hard it's giving me less assistance than pushing it with around 70% of force when sitting. Counteruntuitive. Is this normal for 0.54 firmware?

 

[izeman]

Today I wanted to install BRAKE sensors. I got the ones from pswpower, but they are a bit bulky, and I need to extend the wires to make them fit.
II also have other sensors, CS1-G Magnetic Reed Switch Sensors. They are very small and have a little LED to show their state. They also fit perfectly on my brakes.
But those sensors don't work. They seem to be NC instead of NO. When connected the display shows BRAKE when the magnet is next to the sensor instead of vice versa.
Any idea how those can be used, or if there is a setting in the firmware to change the sensor type?

 

[aleixonunes]

The Ebike open source App started doing the firsts (baby) steps :wink:
In the video you can see the app running in a real Android device with the basics navigation working.
[youtube]oEJaw9lPl_U[/youtube]

 

[skestans]

Man that sucks. Are you using a speed sensor connector, or DuPont wires with heat shrink tube at the ends? If DuPont, maybe you’re not getting a good connection and should try with a speed sensor extension cable so you get the proper connector and good contact.

If your speed sensing works correctly (ie displays the right speed on the display) and you’ve tried so many computers and dongles, the only option left is a bad connection at the plug.

You could try with a Linux live disc and see if stm8flash works there but if it didn’t work on OS X...

I'm using the speedsensor connector with the Dupont wires (4cm long) and heat shrink tubes. And you are right that might be the issue, but they really snuck in tight. So I'm assuming (very bad I know) they work. If I connect the GND, 5V and SWIM and boot the bike via the VLCD5 display then my ST Link V2 boots, so I'm fairly confident (I know very bad) that the connection is working.

I used a speed sensor extension cable at first, but a lot of people point out to the long leads being the problem.

regards,
LVX

I took the motor controller board out today. It did come out easily.


I hooked up with the short Dupont wires (which still doesn't make sense, as the wire from the board up to connector is ~30cm long) and it did not connect. I tried 3 different type of STLink V2 dongles.

I removed some of the coating to measure from the PCB pads. All of the wires gave a good readout. I tried testing what is what, and with some help of google I ended up with (right-to-left):
brown = VCC
black = SWIM
orange = GND
purple = reset


Could it be that this board is somehow protected for readout, or is a protected firmware still communicating?

I just wanted to add that I was doing some debugging to see if I could find out if the board is actually powered by the ST Link.

So I hooked up an UART dongle to read out the main connector, the 9 pin version. Immediately after plugging in the ST Link dongle the UART started to write data to the console, all 16 byte lines of code.


As soon as I click on the STVP "read all tabs" icon, the UART stopped. So it definitely rebooted the MCU.

I then went on to check if I have to correct MCU, which luckily I did.


Can someone tell me if can try out a different type of test to see if the SWIM works?

regards,
LVX

You could figure out which pins are the lines you need directly on the MCU and use pigtails to connect the programmer.

 

[izeman]

You could figure out which pins are the lines you need directly on the MCU and use pigtails to connect the programmer.

Could you please stop QUOTING WHOLE POSTS?

 

[Zelenaar]

The Ebike open source App started doing the firsts (baby) steps :wink:

 

[skestans]

You could figure out which pins are the lines you need directly on the MCU and use pigtails to connect the programmer.

Could you please stop QUOTING WHOLE POSTS?

Someone is fussy today.

Let me tell you about this formidable invention that is the scroll wheel! Any mouse has it nowadays, starting at 2$. I hope you figure it out.

 

[casainho]

You could figure out which pins are the lines you need directly on the MCU and use pigtails to connect the programmer.

Could you please stop QUOTING WHOLE POSTS?

Someone is fussy today.

Let me tell you about this formidable invention that is the scroll wheel! Any mouse has it nowadays, starting at 2$. I hope you figure it out.

If possible, is better to try to keep this clean, much less effort for everyone and so focus on what is really important.

 

[izeman]

Someone is fussy today.
Let me tell you about this formidable invention that is the scroll wheel! Any mouse has it nowadays, starting at 2$. I hope you figure it out.

Just don't do it. No need to quote quotes of quotes just because you are too lazy to delete unneeded stuff.

 

[windburner]

Ok, so if assist level is set too high you get less assist. I've set highest assist level to 0.9 and couldn't get more than 450-500W up a steep hill. Reduced to 0.200 motor is assisting now with 650W max, still not 750W I was getting with 0.19 firmware.

Another really strange thing is that if I really push pedals hard it's giving me less assistance than pushing it with around 70% of force when sitting. Counteruntuitive. Is this normal for 0.54 firmware?

Thanks for continuing to provide insight into this assist strangeness you've encountered with v0.6.7/v0.54. Do I assume correctly that the wattage values you have given are "as reported" thru the SW102 display? Have you compared "as reported" voltage to the reading from a volt meter? (My "as reported" voltage is ~1.5V higher than my trusted meters.)

I am looking forward to a sufficient break in our wintry weather this week in order to get out and do real world testing of this update. I will plan to connect an amp meter in order to get some 'on the fly' measurements of power consumption at various assist level settings (especially values near 0.200 and higher).

Now, the "push harder and get less assist" strangeness should make the dreaded hill climb to get back home a bit more than interesting. :?

 

[hefest]

Ok, so if assist level is set too high you get less assist. I've set highest assist level to 0.9 and couldn't get more than 450-500W up a steep hill. Reduced to 0.200 motor is assisting now with 650W max, still not 750W I was getting with 0.19 firmware.

Another really strange thing is that if I really push pedals hard it's giving me less assistance than pushing it with around 70% of force when sitting. Counteruntuitive. Is this normal for 0.54 firmware?

Thanks for continuing to provide insight into this assist strangeness you've encountered with v0.6.7/v0.54. Do I assume correctly that the wattage values you have given are "as reported" thru the SW102 display? Have you compared "as reported" voltage to the reading from a volt meter? (My "as reported" voltage is ~1.5V higher than my trusted meters.)

Voltage displayed on sw102 is 51.0V. Voltmeter measured 50.7V. So the "real" power (W) is even smaller then displayed.
I'd need to replace battery connector to be able to connect wattmeter.

 

[Powerhour]

So.. since I stupidly modded my shunt, is there anyway to change the resistance value of it in the firmware? Or would the best thing for me to do would be to pick the hotglue off it and soldersuck off what I’ve done?
Thanks

 

[baloubic]

Ok, back with another strange question... I can’t get my speed sensor working.

At first, I just thought it was I had it mounted in the wrong direction (horizontal), and wasn’t getting a reading. Then I turned it, but still nothing. Lastly, I tried connecting a spare sensor from a previous tsdz2; then I got a working speed reading, but only once. After that one first time, it doesn’t matter what speed sensor I connect, I get no reading. I don’t have a light connected to the splitter anymore, and I’ve only switched wires with the battery disconnected.

Can you fry a speed sensor somehow, or what would you suggest I do? I figured it could have something to do with the light switch, so I’ve tried after both long plus press (dim screen, 850c, as when you would start the light) and without. Nothing happens with the speed signal. It’s stuck on 0 regardless if the magnet passes close by in any direction. Does it have to be activated in firmware somehow? Since I got that initial reading, something tells me there’s something I can do to get it working again.

I want to see how fast I go, cause this bike feels so awesome to ride!